The vast majority of biomedical information sources now being published to the Internet employ the World Wide Web. However, it is still extremely difficult in obtaining current, accurate medical knowledge for physicians at a clinical setting or other users seeking for medical information. We propose a novel medical knowledge retrieval system that has the capability in collecting and grouping searches and results through multiple servers and resources. Information discovery by the user will not be challenged by difficulties and differences in query syntax, network addresses across servers and a lack of conceptual organization. The system employs an expressive structured query language and the National Library of Medicine's Unified Medical Language System to provide the necessary terminology in enhancing information retrieval. It will provide the users with a network accessible collection repository for data sharing and reusing. The proposed system will use a cross-platform programming environment-Java to promote wide distribution and implementation.